Indoor localization and mobility management in the emerging heterogeneous wireless networks

(1)

HAL Id: tel-00997657

https://tel.archives-ouvertes.fr/tel-00997657

Submitted on 28 May 2014

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Indoor localization and mobility management in the

emerging heterogeneous wireless networks

Apostolia Papapostolou

To cite this version:

Apostolia Papapostolou. Indoor localization and mobility management in the emerging heterogeneous wireless networks. Architecture, space management. Institut National des Télécommunications, 2011. English. �NNT : 2011TELE0003�. �tel-00997657�

(2)

Thèse présentée pour l’obtention du diplôme de

Docteur de Télécom & Management SudParis

Doctorat conjoint Télécom & Management SudParis et Université Pierre et Marie Curie

Spécialité :

Informatique

Localisation en Intérieur et Gestion de la Mobilité

dans les Réseaux Sans Fils Hétérogènes

Emergents

Par

Apostolia Papapostolou

Soutenue le 31 janvier 2011 à TSP devant le jury composé de :

Mr. André Luc BEYLOT

Professeur à l’université INPT Toulouse Rapporteur

Mr. David SIMPLOT

Professeur à l’Université de Lille1

Rapporteur

Mr. Jean Marie BONNIN

Professeur à Telecom Bretagne

Examinateur

Mr. Marcelo DIAS DE AMORIM Chargé de Recherche CNRS

Examinateur

Mr. Mischa DOHLER

Directeur de Recherche CTTC Barcelone Examinateur

Mr. Sébastien TIXEUIL

Professeur à l’UPMC

Examinateur

Mme Hakima CHAOUCHI

Maître de Conférence-HDR TSP

Directeur de thèse

(3)

Th`

ese de Doctorat

T´

el´

ecom & Management SudParis et

l’universit´

e Pierre et Marie Curie (Paris VI)

Sp´ecialit´e

S

YST`EMES

I

NFORMATIQUES

pr´esent´ee par

Mlle. Apostolia P

APAPOSTOLOU

pour obtenir le grade de

DOCTEUR

conjoint de Télécom & Management SudParis et l’université Pierre et

Marie Curie

Localisation en Int´

erieur et Gestion de la Mobilit´

e

dans les R´

eseaux Sans Fils H´

et´

erog`

enes ´

Emergents

Soutenance pr´evue le 31 janvier 2011 devant le jury compos´e de

Jury

André Luc BEYLOT Rapporteur Prof. à l’université INPT Toulouse

David SIMPLOT-RYL Rapporteur Directeur de Recherche Universit´e Lille1

Mischa DOHLER Examinateur Directeur de recherche CTTC Barcelone

Sébastien TIXEUIL Président du jury Prof. à l’université Pierre et Marie Curie

Marcelo DIAS DE AMORIM Examinateur Charg´e de Recherche CNRS

Jean Marie BONNIN Examinateur Prof. `a Telecom Bretagne

Hakima CHAOUCHI Directrice Maˆıtre de Conf´erence `a Telecom Sudparis

(4)

(5)

Th`

ese de Doctorat

T´

el´

ecom & Management SudParis et

l’universit´

e Pierre et Marie Curie (Paris VI)

Sp´ecialit´e

S

YST`EMES

I

NFORMATIQUES

pr´esent´ee par

Mlle. Apostolia P

APAPOSTOLOU

pour obtenir le grade de

DOCTEUR

conjoint de Télécom & Management SudParis et l’université Pierre et

Marie Curie

Localisation en Int´

erieur et Gestion de la Mobilit´

e

dans les R´

eseaux Sans Fils H´

et´

erog`

enes ´

Emergents

Soutenance pr´evue le 31 janvier 2011 devant le jury compos´e de

Jury

André Luc BEYLOT Rapporteur Prof. à l’université INPT Toulouse

David SIMPLOT-RYL Rapporteur Directeur de Recherche Universit´e Lille1

Mischa DOHLER Examinateur Directeur de recherche CTTC Barcelone

Sébastien TIXEUIL Président du jury Prof. à l’université Pierre et Marie Curie

Marcelo DIAS DE AMORIM Examinateur Charg´e de Recherche CNRS

Jean Marie BONNIN Examinateur Prof. `a Telecom Bretagne

(6)

(7)

Doctor of Science Thesis

Telecom & Management SudParis and

Pierre & Marie Curie University (Paris VI)

Specialization

C

OMPUTER

S

CIENCE

presented by

Miss Apostolia P

APAPOSTOLOU

Submitted in partial satisfaction of the requirements for the degree of

DOCTOR OF SCIENCE

Telecom & Management SudParis and Pierre & Marie Curie University

Indoor Localization and Mobility Management

in the Emerging Heterogeneous Wireless Networks

Commitee in charge:

Andr´e Luc BEYLOT Reviewer Prof. at INPT Toulouse

David SIMPLOT-RYL Reviewer University Lille1 Research Director

Mischa DOHLER Examinator CTTC Barcelone Research Director

S´ebastien TIXEUIL President Prof. at Pierre and Marie Curie University

Marcelo DIAS DE AMORIM Examinator CNRS Researcher

Jean Marie BONNIN Examinateur Prof. at Telecom Bretagne

(8)

(9)

(10)

(11)

R´

esum´

e

Au cours de ces dernières décennies, nous avons été témoins d’une évolution considérable dans l’informatique mobile, réseaux sans fil et des appareils portatifs. Dans les réseaux de communication à venir, les utilisateurs devraient être encore plus mobiles exigeant une connectivité omniprésente à différentes applications qui seront de préférence au courant de leur contexte. Certes, les informations de localisation dans le cadre de leur contexte est d’une importance primordiale à la fois la demande et les perspectives du réseau. De point de vu de l’application ou l’utilisateur, la provision de services peuvent mettre à jour si l’adaptation au contexte de l’utilisateur est activée. Du point de vue du réseau, des fonctionnalités telles que le routage, la gestion de handoff, l’allocation des ressources et d’autres peuvent également bénéficier si l’emplacement de l’utilisateur peuvent être suivis ou même prédit.

Dans ce contexte, nous nous concentrons notre attention sur la localisation à l’intérieur et de la prévision de handoff qui sont des composants indispensables à la réussite ultime de l’ère de la communication omniprésente envisagé. Alors que les systèmes de position-nement en plein air ont déjà prouvé leur potentiel dans un large éventail d’applications commerciales, le chemin vers un système de localisation réussi à l’intérieur est reconnu pour être beaucoup plus difficile, principalement en raison des caractéristiques difficiles liées à l’intérieur et l’exigence d’une plus grande précision. De même, la gestion de handoff dans des réseaux hétérogènes sans fil de futur est beaucoup plus difficile que dans les réseaux traditionnels homogènes. La procédure de handoff doit être transparente pour satisfaire la qualité de service requise par les applications de futur et leurs fonctionnalités, cela ne doit pas dépendre de la caractéristique de l’opération des technologies différentes. En outre, les décisions de handoff devraient être suffisamment souples pour tenir compte aux préférences des utilisateurs d’un large éventail de critères proposés par toutes les technologies.

L’objectif principal de cette thèse est de mettre au point précis, le temps et l’emplace-ment de puissance efficaces et la gestion de handoff afin de mieux satisfaire les applications sensible des utilisateurs en dépendent au contexte dans lequel les utilisateur se trouvent. Pour obtenir une localisation à l’intérieur, le potentiel de réseau sans fil local (WLAN) et Radio Frequency Identification (RFID) comme une technologie autonome pour détection de location sont d’abord ont été étudiés par des expérimentations de plusieurs algorithmes et paramètres dans des plateformes réels ou par de nombreuses simulations, alors que leurs lacunes ont également été identifiés. Leur intégration dans une architecture com-mune est alors proposée afin de combiner leurs principaux avantages et surmonter leurs limitations. La supériorité des performances du système de synergie a été validée par des

(12)

sibilité au contexte peut aussi améliorer la fonctionnalité du réseau. En conséquence, deux types de systèmes qui utilisent l’information obtenue à partir des systèmes de localisation one été proposées. Le premier schéma repose sur un déploiement tag RFID, comme notre architecture de positionnement RFID, et en suivant la scène WLAN analyse du concept de positionnement, prédit l’emplacement réseau de la prochaine couche, c’est à dire le prochain point de fixation sur le réseau. La deuxième méthode repose sur une approche intégrée RFID et réseaux de capteurs / actionneur Network (WSAN) de déploiement pour la localisation physique des utilisateurs et par la suite pour prédire leur prochaine point de handoff aux niveaux des couches de liaison et le réseau. Etre indépendant de la technologie d’accès sans fil sous-jacent, les deux régimes peuvent être facilement mises en oeuvre dans des réseaux hétérogènes.

L’évaluation de la performance démontre les avantages de nos méthodes proposées par rapport aux protocoles standards concernant l’exactitude de prévision, le temps de latence et l’économie d’ énergie. Les mots clés : mobilité, localisation, gestion de handoff, commu-nication des réseaux sans fil, architecture des réseaux hétérogènes, analyse de performance, WLAN, RFID, WSAN.

Mots-cl´

es :

localisation, mobilité, gestion de la handoff, communications sans fil, hétérogénéité, conception d’architecture réseau, analyse de performance, WLAN, RFID, WSAN.

(13)

(14)

(15)

Abstract

Over the last few decades, we have been witnessing a tremendous evolution in mo-bile computing, wireless networking and hand-held devices. In the future communication networks, users are anticipated to become even more mobile demanding for ubiquitous connectivity to different applications which will be preferably aware of their context. Ad-mittedly, location information as part of their context is of paramount importance from both application and network perspectives. From application or user point of view, service provision can upgrade if adaptation to the user’s context is enabled. From network point of view, functionalities such as routing, handoff management, resource allocation and others can also benefit if user’s location can be tracked or even predicted.

Within this context, we focus our attention on indoor localization and handoff predic-tion which are indispensable components towards the ultimate success of the envisioned pervasive communication era. While outdoor positioning systems have already proven their potential in a wide range of commercial applications, the path towards a successful indoor location system is recognized to be much more difficult, mainly due to the harsh indoor characteristics and requirement for higher accuracy. Similarly, handoff management in the future heterogeneous wireless networks is much more challenging than in traditional homogeneous networks. Handoff schemes must be seamless for meeting strict Quality of Service (QoS) requirements of the future applications and functional despite the diversity of operation features of the different technologies. In addition, handoff decisions should be flexible enough to accommodate user preferences from a wide range of criteria offered by all technologies.

The main objective of this thesis is to devise accurate, time and power efficient location and handoff management systems in order to satisfy better context-aware and mobile ap-plications. For indoor localization, the potential of Wireless Local Area Network (WLAN) and Radio Frequency Identification (RFID) technologies as standalone location sensing technologies are first studied by testing several algorithms and metrics in a real experimen-tal testbed or by extensive simulations, while their shortcomings are also identified. Their integration in a common architecture is then proposed in order to combine their key bene-fits and overcome their limitations. The performance superiority of the synergetic system over the stand alone counterparts is validated via extensive analysis.

Regarding the handoﬀ management task, we pinpoint that context awareness can also enhance the network functionality. Consequently, two such schemes which utilize informa-tion obtained from localizainforma-tion systems are proposed. The ﬁrst scheme relies on a RFID tag deployment, alike our RFID positioning architecture, and by following the WLAN scene

(16)

Sensor/Actuator Network (WSAN) deployment for tracking the users’ physical location and subsequently for predicting next their handoﬀ point at both link and network layers. Being independent of the underlying principle wireless access technology, both schemes can be easily implemented in heterogenous networks. Performance evaluation results demonstrate the advantages of the proposed schemes over the standard protocols regarding prediction accuracy, time latency and energy savings.

Key Words:

localization, mobility, handoﬀ management, wireless communications, heterogeneity, network architecture design, performance analysis, WLAN, RFID, WSAN.

(17)

7.4 Theoretical Analysis . . . 131 7.4.1 Time Response . . . 132 7.4.1.1 Standard Protocols . . . 132 7.4.1.2 Scheme A . . . 132 7.4.1.3 Scheme B . . . 135 7.4.2 Energy Consumption . . . 135 7.4.2.1 IEEE 802.11 Scanning . . . 136 7.4.2.2 Scheme A . . . 136 7.4.2.3 Scheme B . . . 136 7.5 Performance Analysis . . . 137 7.5.1 Simulation Setup . . . 137 7.5.2 Accuracy Analysis . . . 139 7.5.3 Energy Consumption . . . 140 7.5.4 Time Latency . . . 140

(21)

Table of contents 15

8 Conclusions 143

8.1 Contributions . . . 144

8.2 Future Directions . . . 146

9 Thesis’ French Version 147 9.1 Les objectifs et Les d´eﬁs . . . 148

9.2 Un Aper¸cu de la th`ese . . . 150

9.3 Le cheminement de travail et les contributions . . . 152

9.4 L’Organisation de la th`ese . . . 155 References 157 List of publications 169 List of acronyms 171 List of figures 173 List of tables 176

(22)

(23)

Chapter 1 Introduction

W

ith the rapid growth of wireless communication and networking technologies, the great advances in mobile computing and handheld devices, and the overwhelming success of Internet, a revolutionary pervasive and mobile communication era is emerging as the natural successor of current mobile communication systems. The goal of this pervasive or ubiquitous computing vision is to create ambient intelligence with core concept the interaction between human with its environment and ultimate goal the enhancement of the user experience from the network. To that aim, an increasingly large numbers of everyday objects scattered throughout the surrounding environment will become smart by having some kind of simple computation and communication technology embedded into them, which will allow them to be connected to each other within local networks and, ultimately, connected to the Internet. Users will become even more mobile demanding to experience unobtrusive connectivity and ubiquitous access to diﬀerent applications anywhere, anytime, by using the best technology from a plethora of interfaces available at the future multi-mode mobile terminals, and without the need for explicit awareness of the underlying communication and computing technology.

For the realization of such ubiquitous environments, location awareness and efficient mobility, in terms of handoff, management are two core concepts. Furthermore, a strong correlation exists between them. The continuous need for determining the unknown loca-tion of an entity stems from its mobility capability. Simultaneously, dealing with issues raised due to mobility can benefit if location information is available. This thesis targets at improving both the localization and handoff processes and proposes taking advantage of the availability of several wireless technologies for tackling more effectively the objectives

(24)

of future communication networks. In the following, the main objectives, challenges and our approaches for achieving that goal are described.

1.1 Objectives and Challenges

Indoor location information is valuable for facilitating the interaction between a user and its environment and consequently the development of location based services (LBS) or more generally speaking context-aware applications where location is a key element of the user’s context. Such applications adapt their functionality depending on the user’s context and they span from applications in users’ everyday life, working environment, commercial and industrial sectors to functions which aim at the performance enhancement of the wireless network functionality. Some typical examples of location-aided applications are:

≤ Ambient Assisted Living: Accurate positioning information is critical for the success of the Ambient Assisted Living (AAL) project [1] which aims at enhancing the everyday life of elderly users and people suﬀering from disabilities, through the use of Information and Communication Technologies (ICT).

≤ Person and Asset Tracking: Tracking of people inside buildings is critical in emergency situations such as fires, earthquakes, or other disasters. Moreover, indoor location systems are useful in hospitals for tracking staﬀ members at any time without their intervention, in museums or schools for keeping track of children location [2]. Tracking of objects or assets is useful for finding the whereabouts of hospital equip-ment in a hospital, finding books inside a library or products inside a warehouse. The location of various physical resources such as printers, projects, and copiers also enables resource discovery applications [3].

≤ Navigation: Indoor location information can be used to build navigating tools in unfamiliar buildings [4], such as airports, train stations, museums, campuses, com-mercial department stores or big oﬃce buildings.

≤ Location-Based Advertising and Social Networking: Location methods can be used for selective and targeted advertising [5] and for providing product information inside retail stores [6].

(25)

1.1 Objectives and Challenges 19

based social networking services such as Facebook, Friendsters, MySpace, etc. by allowing users forming groups based on their social preference and interest.

≤ Network performance improvement User location information can be also ex-ploited to enhance the functionality and the QoS in wireless networks. Such methods have been proposed for location-based access control [7], location-based handoﬀ and in ad hoc networks in order to optimize routing algorithms and network self con-figuration [8]. One step further, combining positioning data with user profiles could significantly facilitate network planning, load balancing, caching of information closer to the user, radio resource management and design of other performance enhancement methods [9].

For the success of the above applications, the design of an accurate and reliable loca-tion determinaloca-tion system is essential. Wireless localizaloca-tion, i.e. localoca-tion estimaloca-tion by using radio signals (RS), has attracted considerable attention in the fields of telecommuni-cation and navigation. The most well known positioning system is the Global Positioning System (GPS) [10], which is satellite-based and is successful for tracking users in outdoor environments. However, the inability of satellite signals to penetrate buildings cause the complete failure of GPS in indoor environments. For indoor location sensing a number of alternative wireless technologies have been proposed, such as infrared (IR), ultrasound, Wireless LAN (WLAN), UltraWideBand (UWB), Radio Frequency Identification (RFID), Bluetooth, wireless sensor networks (WSNs) [11]. However, the indoor radio propagation channel is characterized as site specific, exhibiting severe multipath eﬀects and low proba-bility of line-of-sight (LOS) signal propagation between the transmitter and receiver [12], making accurate indoor positioning very challenging. Moreover, compared to outdoor sys-tems, determining the location of a user or device inside a building is much more diﬃcult not only due to its harsh nature but also due to the requirement of indoor services for higher and more precise accuracy.

Handoﬀ management is the process for keeping active the connection of the mobile user while changing its point of attachment to the network due to mobility. In the future per-vasive networks, several heterogeneous wireless technologies will be available and users will demand ubiquitous access and ”always best” connectivity to a wide range of applications while on the move. For the harmonized integration of these diﬀerent technologies under a common framework, the design of intelligent mobility management schemes is required in order to enable mobile users to experience uninterrupted service continuity anywhere,

(26)

anytime, regardless their underlying access technology. Furthermore, mobility manage-ment schemes should be able to satisfy the requiremanage-ments of emerging applications which are becoming more and more demanding regarding their QoS constraints.

However, the latency during the handoff processes leads to performance degradation. For the case of IEEE 802.11 WLAN wireless access, the handoff process requires from the mobile node to search periodically for better access points to associate with, by scanning all WLAN channels. However this process is power consuming and introduces packet loss, since during scanning the mobile node is not able to be served by its current AP. Mobile IP [13] is a network layer mobility management scheme for IP-based networks. It forwards packets to mobile users that are away from their home networks using IP-in-IP tunnels. Mobile IP handoff is composed of a sequence of stages, one of which includes the detection of a mobile node’s movement to the new network. However, when the mobile node undergoes movement detection, it is unable to receive IP packets, resulting in further performance degradation.

1.2 Thesis Overview

Admittedly, location awareness of users and objects or devices in an indoor environ-ment and their mobility manageenviron-ment across heterogeneous networks are considered as key milestones towards the realization of future mobile communication networks. Furthermore, the strong correlation between these tasks mandates investigating their aspects in parallel, instead of considering them as two independent processes.

This thesis targets the development of location and mobility management schemes with main design goals:

≤ Accuracy. Knowing exactly where someone or something is or moves towards can improve user experience by personalized service delivery and also enhance the network functionality.

≤ Fast time response. Emerging applications will be more demanding in terms of QoS requirements, impelling for fast localization and handoﬀ schemes.

≤ Scalability. The presence of many users should not degrade the system performance. ≤ Generic handoff. The co-existence of heterogeneous networks within which the user

(27)

1.3 Workflow and Contributions 21

≤ Energy-awareness. Since mobile devices are battery constrained, energy consump-tion issues should be taken into account as well.

Future communication systems are envisioned to be heterogeneous offering ubiquitous connectivity, whereby mobile users will be surrounded by diverse but complementary tech-nologies capturing their different needs and requirements. Motivated by this observation, exploring possible synergies and interactions among several technologies was our main ap-proach in order to tackle more effectively our goals.

For indoor location sensing, we focused our attention on two wireless technologies; WLAN and RFID. WLANs, such as IEEE 802.11, is considered as a promising one for oﬀering a low-cost and reliable solution due to its availability in most indoor environments and capability for coordinated communication when in infrastructure mode. However, its accuracy is highly aﬀected in the presence of severe multipath and environmental changes. More recently, RFID has emerged as an attractive technology for accurate location sensing due to the low cost of passive tags, the fast reading of multiple tags, the non Line of Sight (LOS) requirement, the less sensitivity in user orientation. However, the main shortcoming of RFID is considered the interference problem among its components, mainly due to the limited capabilities of the passive tags and the inability for direct communication between readers [14]. In order to overcome the limitations of both technologies we proposed an integration architecture for improving the localization performance.

Regarding the mobility management problem, we focused on the handoff component and we explored the potential of two popular pervasive technologies: RFID and WSAN, for providing fast handoff solution in the case of IP-based mobility over a WLAN access network. However, our proposed schemes can be applied for different link and network level mobility scenarios, making them viable solutions in heterogeneous networks.

1.3 Workflow and Contributions

For achieving the goals and objectives of this thesis the succeeding steps were followed in a chronological order:

≤ Initially, we focused on the case of WLAN fingerprinting positioning approach which is considered as the most popular for low-cost indoor localization. Considering both its deterministic and probabilistic variants, we proposed some simple techniques in order

(28)

to improve the localization accuracy without increasing considerably the complexity and hardware requirements. Reference papers1 include:

-C1 A. Papapostolou and H. Chaouchi, WIFE: Wireless Indoor positioning based on Fingerprint Evaluation, in Proceeding of the 8th IFIP NETWORKING confer-ence, Aachen, Germany, March 2009 [15].

-C2 A. Papapostolou and H. Chaouchi, Orientation - Based Radio Map Extensions for Improving Positioning System Accuracy, in Proceeding of the 6th ACM In-ternational Wireless Communications and Mobile Computing Conference (ACM IWCMC), Leipzig, Germany, June 2009 [16].

-J1 A. Papapostolou and H. Chaouchi, Scene Analysis Indoor Positioning Enhance-ments, Annals of Telecommunications journal, October 2010 [17].

≤ RFID positioning is considered as another attractive solution for location sensing with higher accuracy than WLAN systems. However, not much attention has been given to the collision problem, as far as positioning is concerned, which is the Achilles’ heel of RFID technology. Therefore, we studied extensively the performance of the most popular RFID positioning algorithms in the presence of multiple users. Reference papers include:

-C3 A. Papapostolou and H. Chaouchi, Considerations for RFID-based Indoor Si-multaneous Tracking, in Proceedings of the 2nd Joint IFIP Wireless and Mobile Networking Conference, Gdansk, Polland, September 2009 [18].

-J2 A. Papapostolou and H. Chaouchi, RFID-assisted Indoor Localization and the Impact of Interference on its Performance, in the SI on RFID Technology, Sys-tems, and Applications of the Journal of Network and Computer Applications (Elsevier), April 2010 [19].

≤ Motivated by the benefits but also the limitations of the stand alone solutions, as identified in the previous steps, an integration architecture combining both WLAN and RFID technologies was then proposed. The main idea is to take advantage of the localization accuracy oﬀered by the RFID deployment and the coordination capability of the WLAN infrastructure for minimizing the collision problem on the RFID channel

1_{In the enumeration list, the symbols C, J, B stand for publications in Conferences, Journals and Book}

(29)

1.3 Workflow and Contributions 23

with ultimate goal to enhance the localization accuracy in a time-eﬃcient manner. Reference papers include:

-C4 A. Papapostolou and H. Chaouchi, Simulation-based Analysis for a Heteroge-neous Indoor Localization Scheme, in Proceedings of the 7th IEEE Consumer Communication and Networking Conference (IEEE CCNC), Las Vegas, Nevada, January 2010 [20].

-C5 A. Papapostolou and H. Chaouchi,Exploiting Multi-modality and Diversity for Localization Enhancement: WiFi and RFID usecase, in Proceedings of the 20th IEEE International symposium on Personal Indoor and Mobile Radio Commu-nications (IEEE PIMRC), Tokyo, Japan, September 2009 [21].

≤ Our next step was motivated by the observation that the joint WLAN and RFID architecture could be also used for the purpose of mobility management. Targeting the network layer handoﬀ improvement, we proposed utilizing the RFID deployment in order to minimize the delay of the movement detection phase in the WiFi channel of the IP mobility management process. However, it is worthy mentioning that the proposed architecture is also valid in other mobility networks. Reference papers include:

-C6 A. Papapostolou and H. Chaouchi, RFID-assisted Movement Detection Improve-ment in IP Mobility, in Proceedings of the 3rd IFIP International Conference on New Technologies, Mobility and Security (IFIP NTMS), Cairo, Egypt, De-cember, 2009 [22].

-C7 A. Papapostolou and H. Chaouchi, Handoff Management relying on RFID Tech-nology, in Proceedings of the IEEE Wireless Communication and Networking Conference (IEEE WCNC), Sydney, Australia, April 2010 [23].

≤ In the sequence the unified system architecture for both localization and mobility management was designed and analyzed in:

-B1 A. Papapostolou and H. Chaouchi, RFID Deployment for Location and Mobility Management on the Internet, in H. Chaouchi (ed), The Internet of Things: Connecting Objects, Wiley, John & Sons, May 2010 [24].

(30)

-J3 A. Papapostolou and H. Chaouchi, Integrating RFID and WLAN for Indoor Po-sitioning and IP Movement Detection journal of Wireless Networks (Springer), submitted in November 2009.

≤ Finally, in order to complete the viability of our proposed schemes, we tried to take into account their accompanied energy consumption. Motivated by this, we pro-posed a scheme combining the benefits of RFID and WSAN technologies for handoﬀ improvement with respect to latency and energy consumption. Reference papers include:

-C8 A. Papapostolou and H. Chaouchi, Deploying Wireless Sensor/Actuator Net-works and RFID for Handoff Enhancement in Proceeding of the International Conference on Ambient Systems, Networks and Technologies (ACM ANT), Paris, France, November 2010 [25].

-J4 A. Papapostolou and H. Chaouchi, Handoff Management Schemes in Future Per-vasive Environments, submitted to the journal of Mobile Networks (Springer), SI on Future Internet for Green and Pervasive Media of the journal of Mobile Networks and Applications (Springer), submitted in December 2010.

1.4 Organization of the Thesis

The rest of this thesis is organized as follows. In order to facilitate its presentation, we divided it into two parts: part I is devoted to localization, whereas part II focuses on mobility management aspects. The first chapter of both parts, i.e. chapters 2 and 6, include background and related work essential for the comprehension and highlighting of our contributions. Chapter 3 describes our proposed methods for improving the WLAN fingerprinting localization accuracy, chapter 4 studies the performance of RFID when the collision problem comes into place and chapter 5 ends part I by describing a heterogenous system which combines the benefits of both technologies. Chapter 7 describes and compares the two schemes we propose for mobility management. The first scheme relies on the RFID technology for reducing the network layer handoﬀ latency, while the second scheme utilizes key properties of both RFID and WSAN technologies for handoﬀ latency reduction but also energy saving. Finally chapter 8 summarizes our main conclusions, achievements and open issues for future research.

(31)

Part I

(32)

(33)

Chapter 2 Indoor Localization

L

ocation awareness is one of the main concepts for the realization of ubiquitous and context-aware communications in the envisioned future wireless networks. From ap-plication point of view, location information is essential for enabling Location Based Ser-vices (LBS) in commercial, health-care, public safety and military domains [26]. From network perspective, location awareness can be utilized for enhancing mobility manage-ment functionalities such as routing, mobility prediction, handoﬀ managemanage-ment for quality of service provisioning [7]. Moreover, position information can assist the self-organization, self-conﬁguration of ad-hoc and sensor networks in the future communication networks [8]. Indoor positioning is a complex engineering problem that has been approached by many computing communities: networking, robotics, vision, and signal processing. The wide success and penetration of wireless networks in the realm of consumer applications attracted the attention of most of the research and industry communities for the development of wireless positioning systems, whereby location tracking is achieved with the aid of received radio signal properties. The most well known positioning system is the Global Positioning System (GPS) [10], which is satellite-based and is successful for tracking users in outdoor environments. However, the inability of satellite signals to penetrate buildings cause the complete failure of GPS in indoor environments. Thus, for indoor location sensing a number of alternative wireless technologies have been proposed, such as Cellular, wireless LAN, infrared, ultrasound, ultra-wideband (UWB), RFID, sensor networks [11].

Even though location estimation have been investigated extensively in the last few decades, there is still no absolute solution satisfying all performance requirements. This is because, the indoor radio propagation channel is characterized as site speciﬁc, exhibiting

(34)

se-vere multipath eﬀects and low probability of line-of-sight (LOS) signal propagation between the transmitter and receiver [12], making accurate indoor positioning very challenging.

The aim of this chapter is to provide essential background regarding positioning princi-ples in conjunction with the relevant state-of-the-art. The rest of this chapter is organized as follows: in section 2.1 we first outline the principle aims and requirements a position-ing system should satisfy. Sections 2.2, 2.3 and 2.4 describe the most outstandposition-ing signal metrics, positioning techniques and sensing technologies, respectively, that are employed by the majority of the current state-of-the-art indoor location systems. Finally, section 2.5 provides the chapter summary along with our research directions for indoor positioning.

2.1 Positioning Aims and Requirements

The localization problem is defined as the process of determining the current position of a mobile node or an object within a specific region, indoor or outdoor. The position can be expressed in several ways, depending on the application requirements or the positioning system specifications. For instance, absolute coordinates, relative or symbolic locations are possible formats.

The general framework of a wireless positioning system is illustrated in Figure 2.1. The key concept is to utilize properties of Received Radio Signal (RRS) measurements from several Fixed Reference Points (FRPs) in order to infer the unknown location of the receiver. Initially, radio signals transmitted by the FRPs (such as Access Points or Base Stations) are sensed/measured by the RRS-sensing devices of the receiver and then converted to location-related signal metrics. The reported signal metrics are then processed by the positioning algorithm for estimating the unknown location of the receiver, which is finally utilized by the application. The accuracy of the signal metrics and the complexity of the positioning algorithm define the accuracy of the estimated location.

For evaluating the eﬃciency of localization schemes the major performance objectives are summarized in the following [11],

≤ Accuracy is the most important requirement a location system should satisfy. Its most common metric is the mean distance error, defined as the average of the Euclidean distances between the actual and the estimated locations.

≤ Precision is a metric for evaluating the consistency or reliability of the system over many trials. It is defined as the standard deviation of the distribution of the distance

(35)

2.1 Positioning Aims and Requirements 29

Figure 2.1: Wireless Positioning System. errors or the cumulative probability function.

≤ Complexity involves the computational and communication requirements of the sys-tem. The computational complexity refers to the processing operations of the posi-tioning algorithm, whereas the communication complexity refers to the posiposi-tioning message exchange overhead. These factors mainly aﬀect the time response and/or lifetime of the system.

≤ Scalability refers to the number of nodes that can be simultaneously tracked or the scale of the target area, e.g. metropolitan, campus-wide, within a building, etc. ≤ Deployment and maintenance cost refers to the hardware requirements and labor

eﬀort for installing, setting up and maintaining the positioning system.

≤ Power consumption is also an important aspect especially when the positioning pro-cess is performed by energy-constrained terminals.

≤ Fault tolerance or robustness refer to the ability of the positioning system to perform well even under harsh conditions such as the failure of a positioning component. The wide range of applicability of location information and the diversity of perfor-mance requirements drove the research over several directions, resulting in a vast variety of proposed positioning systems. Consequently, a classification among them is essential for assisting their diﬀerentiation and evaluation. Several classification criteria can be

(36)

ap-plied with the most prevalent ones being the type of metric of the received radio signal measurements, the positioning technique and the wireless sensing technology.

2.2 Received Radio Signal Metrics

The most traditional received radio signal metrics are the Time of Arrival (ToA), Time Diﬀerence of Arrival (TDoA), Received Signal Strength (RSS), Signal to Noise Ratio (SNR), Angle of Arrival (AoA) or combinations of more than one of these metrics [27], [28], [29]. In the following, a more detailed description of these most commonly used signal metrics is provided, followed by a discussion comparing their main advantages and limitations.

2.2.1 Time of Arrival (ToA)

Times of Arrival (ToA) refers to the time needed for a signal to travel from one node to another which indicates their separation distance. Thus, in TOA-based systems, the propagation time of a signal from a FRP to the target node is measured in order to calculate their distance. In order to measure the ToA parameter for a signal traveling between two nodes, these nodes must be timely synchronized and the transmitted signal must be timestamped.

Let s(t) denote the signal transmitted from a node to another at time t. Then, the received signal is expressed as

r(t) = s(t τ ) + n(t), (2.1)

where τ represents the ToA and n(t) is white Gaussian noise with zero mean and a spectral density of No/2.

The most well-known methods for performing ToA estimations are considered to be the correlator or Matched Filter (MF) receivers [30]. According to the correlator-based approach the received signal is correlated with a local template s(t ˆτ ) for various delays ˆ

τ in order to calculate the delay corresponding to the correlation peak. Similarly, the MF approach employs a ﬁlter that is matched to the transmitted signal and estimates the instant at which the ﬁlter output reaches its higher value. Both approaches are optimal in the maximum likelihood (ML) sense for the signal model in eq. (2.1).

However, in real environments, signal distortion due to multipath characteristics, aﬀect the optimality of these conventional schemes.

(37)

2.2 Received Radio Signal Metrics 31

2.2.2 Time Diﬀerence of Arrival (TDoA)

The TDoA approach is followed to replace the absolute synchronization requirement between the target node and the FRPs with the more moderate requirement for relative clock synchronization among the FRPs only. TDoA estimates the time diﬀerence between the arrival times of two signals traveling between the target node and two FRPs, which determines the position of the target node on a hyperbola, with foci at the two FRPs.

One approach for estimating TDoA is to first estimate ToA for each signal traveling between the target node and a FRP, and then to subtract the two estimates. Since the target node and the reference nodes are not synchronized, the ToA estimates include a timing oﬀset, which is the same in all estimates as the reference points are synchronized, in addition to the time of flight. Therefore, the TDoA estimate can be obtained as

τT DoA= ˆτ1 τˆ2, (2.2)

where ˆτi, for i = 1, 2, denotes the ToA estimate for the signal traveling between the target

node and the ith reference point.

Another approach for TDOA estimation is to perform cross-correlations of the two signals traveling between the target node and the FRPs, and to calculate the delay cor-responding to the largest cross-correlation value. The cross-correlation function of these signals is given by integrating the lag product of two received signals over a time period T

ˆ R1,2(τ ) = 1 T r1(t)r2(t + τ )dt. (2.3)

The TDOA is the value τ that maximizes R1,2(τ ), i.e. the range diﬀerences.

2.2.3 Angle of Arrival (AoA)

In AoA-based systems the position is calculated via goniometry. The location of the target node lies on the intersection of several pairs of angle direction lines, each formed by the circular radius from a FRP to the target node. The angle between two nodes can be determined by estimating the AoA parameter of a signal traveling between the nodes. AoA estimates are obtained with the aid of directional antennas based on the beamforming technique [31] or with the aid of antenna arrays based on the principle that diﬀerences in arrival times of an incoming signal at diﬀerent antenna elements include the angle information if the array geometry is known [32].

(38)

For narrow-band signals, time diﬀerences can be represented as phase shifts. Therefore, the combinations of the phase shifted versions of received signals at diﬀerent array elements can be tested in order to estimate the AOA. However, for wide-band systems, time delayed versions of received signals should be considered, since a time delay cannot be represented by a unique phase value for a wide-band signal.

2.2.4 Received Signal Strength (RSS)

Signal strength refers to the power or energy of the signal traveling between two nodes. RSS systems are based on propagation-loss equations which are indicative of the distance r between those nodes, due to signal attenuation as their distance increases. The free-space transmission loss, LB, for instance, is proportional to 1/r2. Thus, if RSS measurement

is used in combination with a path-loss and shadowing model a distance estimate can be obtained. In two-dimensional space, such signal measurement determines the location of one node on the circle centered at the other node with radius their estimated distance.

However, in practice, a signal traveling from one node to another experiences fast (multi-path) fading, shadowing and path-loss, resulting in non-deterministic radio propagation models [33] P L(d) = P L(do) + 10n log d do + Xσ, (2.4)

where P L(d) the path loss for distance d between a FRP and the target node, P L(do) the

free space path loss at reference distance do, n the path loss exponent whose value depends

on the frequency used, the surroundings and building type, and Xσ is a zero-mean Gaussian

random variable in dB having a standard deviation of σdB. The variable Xσ is called the

shadow fading and is used to model the random nature of indoor signal propagation due to the eﬀect of various environmental factors such as multipath, obstruction, orientation, etc. This path loss model is used for calculating the distance d from each RP, based on its transmit power Pt, i.e. RSS(d) = Pt P L(d). Note that this model can be used in both

line-of-sight (LOS) and non line-of-sight (NLOS) scenarios with an appropriate choice of channel parameters.

Some techniques measure SNR ratios although RSS is a stronger function of location as SNR is aﬀected by random fluctuations in the noise process [34].

(39)

2.2 Received Radio Signal Metrics 33

2.2.5 Range-free Metrics

All the above metrics belong to the category of range-based measurements. However, there is another type of metrics which do not rely on such ranging measurements and they are known as range-free or anchor-based metrics [35]. Usually, a heterogeneous network is considered which consists of two types of nodes: (i) anchors which are powerful nodes with established location information and beacon their position to their neighbors, and (ii) blind nodes which listen to such beacons in order to calculate their position.

In [36] the simple centroid algorithm is proposed for estimating the unknown location of a blind node based on the announced locations of the anchor nodes. In DV-HOP [37] anchors, instead of single hop broadcasts, ﬂood their location throughout the network maintaining a running hop-count at each node along the way. Blind nodes calculate their position based on the received anchor locations, the hop-count from the corresponding anchor, and the average-distance per hop; a value obtained through anchor communication. Like DV-Hop, the Amorphous positioning algorithm proposed in [38] uses oﬄine hop-distance estimations, improving location estimates through neighbor information exchange. In [35], an area-based scheme, called APIT, performs accurately location estimations with the irregular radio pattern and random node placement. The main idea of APIT is to divide the whole network into triangular regions among anchors, and then to determine the possible position of a blind node via the aggregation of the two distinct triangular regions. Consequently, the position of the node can be estimated by calculating the Center of Gravity (CoG) of the intersections of the triangles where the node resides.

2.2.6 Comparison

Both ToA and TDoA metrics require strict time synchronization, either between both target nodes and the FRPs or between the FRPs only, respectively. Thus, such metrics are most suited for cellular networks since the receiving nodes are typically synchronized to base stations.

Obtaining AoA measurements is more expensive in implementation compared to ToA and TDoA due to the utilization of special hardware such as antenna arrays, and complex transmission techniques such as beamforming. Moreover, it requires a minimum distance between the receivers which results in additional costs and larger node sizes. Furthermore, this technique is highly sensitive to multipath, NLOS conditions, and array precision.

(40)

mea-surements for most of the receivers. Thus, obtaining RSS information is much simpler than applying signal processing techniques to extract the time or angle of arrival. Since RSS po-sitioning is based on theoretical or empirical models in order to convert the received signal strength measurements to distance estimates its performance depends highly on the chan-nel behavior and the accuracy of the employed radio propagation model. However, node mobility and unpredictable variations in channel behavior, which are even more intense in a complicated indoor space, can occasionally lead to large errors in distance evaluation. Also, this technique is very susceptible to noise and interference.

Finally, comparing range-based and range-free metrics, one would say that selecting the optimal one depends on the assumptions for the network. For instance, range-free metrics are considered as cost-eﬀective solution and thus, more suitable for wireless sensor networks where sensor nodes have limited hardware capabilities. The performance of range-free metrics depends on the density of the anchor nodes and the complexity of the positioning algorithm.

2.3 Principle Location Estimation Techniques

Indoor location systems can be classiﬁed based on the principle approach followed by the positioning algorithm. [39] and [28] provide interesting surveys on the basic positioning techniques and taxonomies of localization systems based on them. There are three prin-ciple classes of positioning techniques, namely proximity, triangulation, either lateration or angulation, and scene analysis, which are employed either alone or in combination and in either their baseline or enhanced version by any location determination system. In the sequence, we describe the general mechanism, performance advantages and limitations of each approach.

2.3.1 Proximity

Proximity-based localization approaches provide symbolic relative location information and their key concept is the ”nearness” to objects with known positions, as shown in Figure 2.2(a). Usually, proximity-based algorithms rely on a dense grid of antennas, each having a well-known position and employ range-free signal metrics where location-aware objects are considered as the anchor nodes. The identiﬁcation of such objects such as credit card point of cell-ID, Cell of Origin (CoO) [40], topology or connectivity information [41] and physical

(41)

2.3 Principle Location Estimation Techniques 35

(a) P roximity (b) A ngulation (c) L ateration

Figure 2.2: Principle Positioning Techniques.

contact detection with the aid of sensors [42] are examples of proximity-based approaches that can be employed for retrieving location information.

In general, proximity methods are considered as simple techniques but with limited capabilities regarding their accuracy performance. For enhancing their accuracy, hardware-based solutions, such as a denser deployment of sensors or identiﬁable objects are required, resulting in higher cost for the development and maintenance of the positioning system.

2.3.2 Triangulation

Triangulation uses the geometric properties of triangles to estimate the target location. It has two derivations: lateration and angulation which diﬀer in the method for obtaining range estimations between the target node and each of the ﬁxed reference points.

2.3.2.1 Lateration

Lateration or distance estimation techniques determine the position of an object by measuring its distance from several ﬁxed reference points. For a two dimensional space at least three such reference points are required which do not lie in the same line. In Figure 2.2(c), the estimated location is the intersection point of the cycles with centers the reference points and radius the corresponding estimated distances from each one of them. Lateration methods employ the range-based signal measurements of RSS, ToA, TDoA or combinations, and can be further classiﬁed accordingly.

Lateration systems require coverage from at least three reference points in order to provide a reliable location estimate. Moreover, the performance of a speciﬁc lateration ap-proach shares the advantages and disadvantages of the corresponding signal measurements,

(42)

i.e. RSS, ToA or TDoA metrics.

2.3.2.2 Angulation

Angulation techniques are very similar to lateration methods, with the diﬀerence that angles instead of distances are measured. For a two-dimensional space, two at least reference points are required which are equipped with directional antennas or support advanced transmission techniques such as beamforming. Based on the received angle of arrival (AoA) measurement of both transmitted signals, the known distance between the reference points and triangle properties, the unknown position of the receiver is calculated. Figure 2.2(b) illustrates this concept.

Angulation-based approaches are very accurate and precise. However, their dependency on advanced hardware and transmission techniques increases the cost and complexity of the positioning system and limits their adoption and incorporation for a low cost, simple and fast positioning solution.

2.3.3 Scene Analysis

The main concept of scene analysis methods, also known as fingerprinting, is that special features of the scene observed at a speciﬁc position are exploited for describing and subsequently identifying that position. Thus, the location of an unknown point can be inferred based on the similarity of such observed scene characteristics.

Figure 2.3 depicts the general mechanism of scene analysis localization. Such methods require an offline phase for learning the radio characteristics in a specific area under study. Such radio characteristics may correspond to any radio signal metrics, either range-based or range-free; RSS is though the commonly selected metric. This signal information is then stored in a database called Radio Map. During the online localization phase, the receiver’s unknown location is inferred based on the similarity between the Radio Map entries and the real-time RSS measurements. The similarity in signal space can be based either on pattern matching techniques (deterministic schemes) or on probability distributions (probabilistic schemes). The type of these features and the way they are represented define the accuracy and complexity of this positioning method. In general, deterministic scene analysis com-pared to the probabilistic case is simpler but less accurate way for discriminating among different area positions.

(43)

2.4 Popular Location Sensing Technologies 37

Figure 2.3: Scene Analysis Positioning Technique.

or speciﬁc hardware. Additionally, they are based on the passive observation of features which do not correspond to geometric distances or angles, the measurement of which com-promises more power consumption. The main disadvantages of such methods are the requirement of a pre-phase for capturing these features and their higher dependency on en-vironmental changes, which cause inconsistency of the signal behavior between the training phase and the time of the actual location determination phase.

2.4 Popular Location Sensing Technologies

All aforementioned positioning techniques can be based on any available technology by taking advantage of the characteristics of the corresponding emitted signals [11]. The most common technology types applied for positioning technologies are outlined in this section. Satellite signals are very successful for outdoor positioning with GPS [10] being the most famous system employing them. However, their inability to penetrate inside buildings cause their complete failure for indoor positioning and therefore are excluded from our relevant survey. This section also provides the current state-of-the art in location systems, classiﬁed

(44)

based on their employed wireless technology. [11], [28] and [39] provide a more extensive survey of location systems.

2.4.1 Infrared (IR)

IR technology is attractive for indoor positioning systems because it is available on board of various, very common wired and wireless devices, such as TV, printer, mobile phones, PDAs, etc. The Active Badge location system [41] developed at AT & T Cambridge in 1990s is one of the ﬁrst indoor badge sensing systems. It uses diﬀuse infrared technology and follows the proximity based positioning approach. Each person to be located wears an active badge which emits on demand a globally unique infrared signal every 10 seconds. Within each target area such as a room, one or more pre-build infrared sensors detect the emitted signals from the badges and forward this information to a central server to which they are connected via wired links. The server upon reception of this data, estimates the position of the detected badges and provides it to several location-aware applications.

However, an IR-based positioning system, which offers absolute position estimations, needs LOS communication between transmitters and receivers without strong light in-terference. Thus, the coverage range per infrastructure device is limited within a room. Furthermore, an IR signal is influenced by fluorescent light and sunlight. Finally, the wired links for connecting the sensors increase its deployment cost.

2.4.2 Ultrasound

Ultrasound positioning systems provide a kind of inexpensive positioning solutions. Usually the ultrasound signals used to locate objects need to be combined with RF sig-nals, which perform synchronization and coordination in the system. These ultrasound positioning systems increase the system coverage area.

The Active Bat positioning system [43] also developed at AT & T Cambridge uses ultrasonic technology and follows the ToA-based lateration technique for determining the position of an active bat, which is a tag carried by a person or attached to an object. Sensor nodes are mounted on the ceiling of the target area in a grid fashion. A controller sends requests via short range to the bats and simultaneously a synchronized reset signal to the ceiling sensors using a wired serial network. In response to the request packets sent by the controller, each bat broadcasts a pulse of ultrasonic to the grid of the ceiling sensors. Each ceiling sensor measures the time interval from reset to ultrasonic pulse arrival and computes

(45)

its distance from the Bat. The local controller then forwards the distance measurements to a central controller, which performs the lateration computation. The location estimation of the Active Bat is more accurate than Active Badge [41] and it can also provide orientation information. However, the performance of ultrasonic is inﬂuenced by the reﬂection and obstacles between tags and receivers, which degrades the system accuracy. Finally, the ceiling sensors need to be connected through cables which increases the deployment cost of the system.

The Cricket Location Support System [44] also uses ultrasound emitters as infras-tructure and follows the ToA-based lateration. However, the computations are performed locally by the objects to be located for reducing the cost and ensuring more privacy. The emitters also transmit RF signals for synchronization of the ToA measurements and for-warding their location information in a decentralized fashion. Such location information is used for proximity based positioning in case of the failure of the lateration due to insuﬃcient number of received ultrasonic beacons.

2.4.3 Cellular network

Indoor positioning based on mobile cellular network is possible if the building is covered by several base stations or one base station with strong RSS received by indoor mobile clients. Otsasen et al. presented a GSM-based indoor localization system in [45]. Their key idea that makes accurate GSM-based indoor localization possible is the use of wide signal-strength fingerprints. The wide fingerprint includes the six strongest GSM cells and readings of up to 29 additional GSM channels, most of which are strong enough to be detected but too weak to be used for efficient communication. The higher dimensionality introduced by the additional channel dramatically increases localization accuracy. They present results for experiments conducted on signal-strength fingerprints collected from three multi-floor buildings. The results show that their indoor localization system can differentiate between floors and achieve median within-floor accuracy as low as 2.5 m.

2.4.4 Wireless Local Area Network (WLAN)

WLAN-based indoor positioning is an example of low cost positioning technology. It uses the existing infrastructures in indoor environments since the 802.11 wireless technology is inexpensive and widely deployed on campuses, hospitals, airports, commercial environ-ments etc. However, the accuracy of location estimations based on the WLAN signals is

(46)

aﬀected by the complex behavior of signal propagation [46], and by various elements in indoor environments such as movement and orientation of human body, the overlapping of APs, the nearby tracked mobile devices, walls, doors, etc.

The Daedalus project [47] developed a WLAN proximity-based system for coarse-grained user location. A mobile host estimates its location to be the same as the access point to which it is attached. Therefore, the accuracy of the system is limited by the access point density.

RADAR [48] proposed by Microsoft Research group is a deterministic fingerprinting system which uses RSS measurements from the existing WLAN. During an offline phase, the system builds a radio map for the RF signal strength from a fixed number of APs, either by calibrating the area or by applying a radio propagation model. During normal operation, the RF signal strength of the mobile client is measured by a set of fixed APs and is sent to a central controller. The central controller uses a k-Nearest Neighbor (k-NN) approach to determine the location from the radio map that best fits the collected signal strength information.

The Aura system proposed in [49] uses two techniques: Pattern Matching (PM) and Triangulation, Mapping and Interpolation (TMI). The PM approach is very similar to the RADAR approach. In the TMI technique, the physical position of all the access points in the area needs to be known and a function is also required to map signal strength onto distances. Based on this information, a set of training points at each trained position is generated. The interpolation of the training data allows the algorithm to use less training data than the PM approach. During the online phase, they use the approximate function they got from the training data to generate contours and they calculate the intersection between diﬀerent contours yielding the signal space position of the user. The nearest set of mappings from the signal-space to the physical space is found by applying a weighted average, based on proximity, to the signal space position.

The Horus [50] is WLAN RSS ﬁngerprinting localization system which deﬁnes the pos-sible causes of variations in the received signal strength vector and devises techniques to overcome them, namely providing the correlation modeler, correlation handler, continuous space estimator, and small-space compensator modules. Moreover, it reduces the computa-tional requirements of the location determination algorithm by applying location-clustering techniques.

The Nibble location system, from UCLA, is a WLAN-based scene analysis scheme which uses a Bayesian network to infer a user location [51]. Their Bayesian network model includes

(47)

nodes to be localized, noise, and access points (sensors). The signal to noise ratio observed from an access point at a given location is taken as an indication of that location.

Ekahau [52] is a commercial real-time location system (RTLS) which also uses the WLAN infrastructure but combines site calibration and the RSS-based triangulation tech-nique for determining the location of WiFi enabled devices.

2.4.5 Bluetooth

Bluetooth is a low-cost and low-power technology and many devices are already equipped with it. Thus, it can be used in location sensing. However, a disadvantage of Bluetooth-based positioning system is that the system can only provide accuracy from 2 to 3 m with the delay of about 20s. Furthermore, Bluetooth positioning systems suﬀer from the draw-backs of RF positioning technique in the complex and changing indoor situations. [53] is a bluetooth-based positioning system.

2.4.6 Radio Frequency Identiﬁcation (RFID)

RFID is a means of storing and retrieving data through electromagnetic transmission to an RF compatible integrated circuit. Thus, it is not only for the indoor positioning applications, but also provides many potential services for the demands of users. The advantage of an RFID positioning system is that cheap, light and small tags can be taken by people to be tracked. The RFID system can uniquely identify equipments and persons tracked in the system. However, the proximity and absolute positioning techniques need numerous infrastructure components installed and maintained in the working area of an RFID positioning system.

SpotON [54] is RFID positioning system which uses RSS measurements to estimate the distance between a target tag and at least three readers and then applies trilateration on the estimated distances.

LANDMARC [55] employs also the RFID technology but follows a scene analysis ap-proach by using readers with diﬀerent power levels and reference tags placed at ﬁxed, known locations as landmarks. Readers vary their read range to perform RSS measurements for all reference tags and for the target tag. The k nearest reference tags are then selected and their positions are averaged to estimate the location of the target tag.

WhereNet positioning system [56] oﬀered by Zebra Technology company is another commercial RTLS based on RFID and follows a sophisticated TDOA algorithm for locating